Objective: To determine the prevalence of unverifiable ("ghost") publications in applications to an otolaryngology residency program through the Electronic Residency Application Service (ERAS), correlate with applicant characteristics, and determine if incidence changed after the addition of PubMed (PMID) numbers in 2008.
Study design and setting: Cross-sectional study of residency applications before and after inclusion of PMID numbers at an academic otolaryngology program.
Subjects and methods: Applications for 2007 and 2008 were reviewed. Publications were verified against Medline, Google Scholar, PubMed, ISI Web of Science, and Google. Ghost publications were defined as journals, books, abstracts, or posters that could not be verified as presented, published, or including the applicant author.
Results: In total, 489 applications were reviewed: 243 before PMID numbers were requested and 246 after. Of 2300 listed publications, 125 (5%) were not actual publications and 460 (20%) were in pending status. Forty-five percent (775/1715) could not be verified: 660 of 953 (69%) abstracts/posters, 18 of 47 (38%) chapters, and 97 of 715 (14%) journal articles. Abstracts/posters and book chapters were hardest to verify. The proportion of overall reported publications that could be verified was lower following the addition of PMID to the ERAS application (P = .0003), and the proportion of verifiable journal articles was unchanged from 86.0% to 86.9% (P = .62). Unlike previous findings, gender and medical school ranking were not associated with ghost publications.
Conclusion: A substantial number of publications, especially book chapters and posters/abstracts, listed on otolaryngology residency applications could not be verified. The addition of the PMID to applications did not reduce the number of ghost journal publications.